The present invention disclosed herein relates to an impedance matching, and more particularly, to an impedance matching method used in an electric system such as a plasma system, a nuclear magnetic resonance system, a communication system, and a power transmission line, and a system performing the same.
If impedance is mismatched between a power source and a load, supplied power to the load cannot be maximized, and moreover controlling of the supplied power is not precisely carried out. Accordingly, an electric system, such as a plasma system, a nuclear magnetic resonance system, a communication system, a high frequency induction heating device, and a power transmission line, includes an impedance matching network between a power source and a load, in order to overcome the above mismatching impedance. For example, a plasma chamber system for fabricating a semiconductor device includes an RF electrode connected to an RF power source and an impedance matching network between the RF power source and the RF electrode. General contents for impedance matching are disclosed in “Microwave Engineering” (Addison-Wesley publisher) in page. 281-328 written by David M. Pozar, and also techniques relating to impedance matching of a plasma system are disclosed in U.S. Pat. No. 3,569,777, U.S. Pat. No. 4,112,395, U.S. Pat. No. 4,557,819, U.S. Pat. No. 5,187,454, U.S. Pat. No. 5,585,766, U.S. Pat. No. 5,621,331, and U.S. Pat. No. 5,689,215, and International Publication No. WO9724748.
FIG. 1 is a flowchart for illustrating a typical impedance matching method.
Referring to FIG. 1, the typical impedance matching method measures electrical characteristics (e.g., current, voltage, and phase) in operation S1, control parameters are extracted from the measured electrical characteristics of a power transmission line to control a matching network in operation S2, and then the matching network is controlled through the extracted control parameters in operation S3.
According to typical methods, a matching network generally includes a plurality of variable capacitors of which capacitances can be controlled through operations of control motors. The extracting of the control parameter in operation S2 includes extracting information for an impedance magnitude and a phase of the power transmission line from a measured current, voltage, and phase difference of the power transmission line, and then by means of the information, capacitances of the variable capacitors required for impedance matching can be calculated.
However, the typical methods cause various technical limitations as follows:
(1) convergence failure for a matching state according to brute dependency on an initial state;
(2) matching delay according to instability near a matching position; and
(3) haunting issue according to high dependency on a load and power transmission line impedance.
That is, the above mentioned limitations may not be overcome by a typical matching method because a capacitance magnitude required for matching is determined based on an un-normalized impedance magnitude. This will be described later in more detail.
FIGS. 2 and 3 are diagrams of brute dependency on an initial state in typical impedance matching.
As described above, a typical impedance matching method includes calculating required capacitances based on a magnitude and a phase of measured impedance. At this point, FIGS. 2 and 3 are diagrams for illustrating a mapping method in a capacitance space used for calculation according to a typical matching method. More specifically, FIGS. 2 and 3 are diagrams for illustrating an impedance magnitude and an impedance phase in a power transmission line, respectively, in a capacitance space expressed by coordinates C1 and C2. At this point, FIGS. 2 and 3 illustrate simulation results when an impedance of a load is 5+50j. C1 and C2 represent capacitances of the variable capacitors, respectively.
On the other hand, the solid line of FIG. 2 represents a contour line that connects points of 50 ohm impedance required for matching, and the solid line of FIG. 3 represents a contour line that connects points of a 0° phase required for matching. Accordingly, point corresponding to matching state is point where the solid lines of FIGS. 2 and 3 intersect, and the matching point is illustrated as a small rectangle in FIGS. 2 and 3. On the other hand, a position corresponding to an asterisk represents an initial state.
According to a typical matching method, variations (i.e., ΔC1 and ΔC2) of the variable capacitors are respectively determined by the magnitude and phase of the measured impedance. That is, as illustrated in FIG. 2, when the measured impedance is more than 50 ohm expressed by an arrow A1, a corresponding motor is driven to increase C1, and when the measured impedance is less than 50 ohm expressed by an arrow A2 or an arrow A3, a corresponding motor is driven to decrease C1. According to this matching method, points corresponding to arrows A1 and A2 approach the matching state, a point corresponding to arrow A3 is father away from the matching state. That is, according to the typical matching method, matching convergence is dependent on an initial position of a capacitance space because there may be a region (e.g., FR) not converged into a matching state. A diverging issue of matching may identically occur in an impedance phase. Especially, as illustrated in FIG. 2, the matching failure region FR may become excessively broader in a typical method.
Additionally, there are two points of 50 ohm impedance magnitude required for matching with respect to one coordinate C1 in a region R1 of FIG. 2. Like this, when a capacitance space is used, the coordinate C1 may not be to a point of 50 ohm impedance magnitude require for matching in one-to-one correspondence. Therefore, there may be ambiguity in determining a direction of a matching trajectory. This ambiguity may be another reason causing the divergence of matching.
FIG. 4 is a diagram for illustrating instability around a matching point according to typical impedance matching.
Referring to FIG. 4, as a matching trajectory approaches toward a matching state, variations of variable capacitors required for impedance matching are delicately controlled. However, because the typical matching method determines the direction and speed of a matching trajectory based on un-normalized magnitude and phase of impedance, it is difficult to precisely control C1 and C2 around the matching point. As a result, according to the typical matching method, as illustrated in FIG. 4, a spiral matching trajectory may occur, which delays a time for reaching impedance matching. If a gradient of impedance in a capacitance space, which will be described below, is large, the time delay for impedance matching is greatly increased.
FIG. 5 is a diagram for illustrating high dependency on load impedance in typical impedance matching. In more detail, FIG. 5 illustrates a simulation result for a magnitude of impedance when impedance of a load is 1+50j.
Comparing FIG. 5 having a load impedance of 1+50j with FIG. 2 having a load impedance of 5+50j, as a real part in the load impedance is decreased, a matching contour line (a contour line of 50 ohm impedance magnitude required for matching) is shrunken to a narrow region of a capacitance space, and an gradient of a impedance magnitude is drastically increased around the matching contour line. This increase of the gradient value makes a precise capacitance control around the matching point more difficult. Consequently, as illustrated in FIG. 6, a matching trajectory may abnormally move around the matching point (i.e., the haunting issue).
Another impedance matching method in a continuous wave (CW) mode will be described. That is, impedance is measured or calculated at a measurement point. If impedance of a measurement point and reactance of variable devices in a matching system are known, load impedance can be calculated. Therefore, a matching condition can be found through the above load impedance and thus control parameters can be determined.
An RF power source operates as a pulse mode, which is periodically turned on or turned off. In this case, after a matching condition is found by operating an RF power source temporarily in a CW mode, it is typically applied to a pulse mode. However, if a load changes as time elapses, the above typical technique cannot completely remove a reflected wave. Accordingly, the typical CW mode matching method needs to be modified in a pulse mode method or requires a new method in order to deal with pulse matching.